Global distribution of methane emissions: an inverse analysis of 2019 observations from TROPOMI

Abstract:

The recent launch of TROPOMI offers unprecedent opportunity to more accurately quantify global methane sources from a top-down perspective. We evaluate the TROPOMI methane retrievals with GOSAT and TCCON measurements using GEOS-Chem as an inter-comparison platform. After eliminating the discrepancies from a priori and averaging kernel, TROPOMI data is globally unbiased between 60°S and 60°N compared to GOSAT and TCCON, with a systematic error (regional bias) of 13 ppbv. TROPOMI enables us to perform global inversion of the sources and sinks of methane at 2°x2.5° resolution and using only one year of satellite measurements. We use it for optimal estimation of monthly wetland emissions, annual non-wetland emissions, and annual hemispheric OH concentrations in 2019. Prior information includes a new bottom-up inventory of emissions from fuel exploitation based on national reports to the United Nations Framework Convention on Climate Change (UNFCCC). Our analytical inversion quantifies the information content from TROPOMI observations and enables an ensemble of inversions to characterize optimization errors. Results are compared to a 7-year inversion of GOSAT data using the same inversion system. The TROPOMI 1-year inversion provides comparable information content with the 7-year GOSAT inversion for anthropogenic emissions, but less for wetlands, in part because cloudiness limits TROPOMI observations over tropical wetlands. We find underestimates of oil and gas emissions in Texas and Middle East and overestimates of coal mining emission in China in the national inventories reported to the UNFCCC, consistent with the inversion results from GOSAT. Both posterior emissions also show downward adjustments in Australia and Europe. Our optimized global methane source for 2019 is 545 Tg a^-1, with a methane lifetime against oxidation by tropospheric OH of 9.1 years.